Thermal balance prediction method based on genetic algorithm and LSTM

Traditional thermal balance relies on artificial experience or simple control algorithms, which is difficult to adapt to complex heating networks and dynamic user needs. In this paper, a long-term and short-term memory network (LSTM) method based on genetic algorithm (GA) optimization is proposed. By analyzing the historical operation data such as temperature, flow and pressure of the heating system, the opening degree of each branch valve in the heating system is predicted to realize the optimal control of thermal balance. It is verified by experiments that the use of this method in central heating projects can alleviate the problem of thermal imbalance, and provide theoretical support for the realization of efficient, energy-saving and sustainable heating systems.